Abstract: The minimum void ratio is one of the most fundamental physical indexes of coarse-grained soil; but up to now, its direct prediction method is lacking for common or oversize coarse-grained soils with different gradations, and its calculation relies on laboratory and field tests. This paper presents a particle packing model for the direct prediction and verifies it through testing on the soil ideal sphere particles. Considering the complicated particle shapes of natural and engineering coarse-grained soils, linkages between the model parameters and particle shapes are needed; then we adopt a hypothesis of a power function relationship of the minimum void ratio versus particle size that holds true for monosize soils. We calibrate the parameters using an optimization algorithm and the experimental data of void ratio tests, and predict the minimum void ratio for different gradations. Application to several practical projects shows satisfactory results of this new method in predicting the minimum void ratio of common and oversize coarse-grained soils and its great value in engineering application.

Key words: particle packing model, parameter optimization algorithm, coarse-grained soil, oversize coarse-grained soil, minimum void ratio